Buried Soils

"In the Pinnacles area of Badlands National Park, there are at least 87 successive fossil soils in 143 m of stratigraphic section." — Greg Retallack, GSA Bulletin, 1983¹

"Clearly Meert considers that paleosols have the potential to refute the global Flood.  We agree!" — Tas Walker, Journal of Creation, 2003²

"The nodules truncate and are truncated by the burrow walls indicating that the burrows are part of the original soil." — Greg Retallack on an Ordovician paleosol in Pennsylvania, 1992³

Summary

A paleosol is a buried soil.  Not a layer of mud that looks soily, but a bed with the three things a soil has and a sediment does not:  root traces that taper and branch downward, horizons whose boundaries grade instead of snapping, and a blocky or prismatic internal structure that has destroyed the parent bedding.

Those criteria are not a mainstream invention that young-earth geologists reject.  Tas Walker copied all three out of Retallack's textbook into Journal of Creation in 2003, and said in the same paper that paleosols could refute a global Flood.²²  There is no argument about the test.  The argument is about whether particular beds pass it, and how long each one took.

Ten conclusions, stated first.

1. The young-earth complaint about the chronometers is correct, and should be printed before a critic prints it. The two published equations that turn a soil profile into years rest on ten and nine data points. Nathan Sheldon and Neil Tabor, whose review reproduces them, say the clay-horizon function "falls apart for very long-formation times" and that the carbonate-nodule function has an R² of 0.57 and a standard error of 1.8 thousand years.⁴⁴ Mary Kraus relays Wright's 1990 verdict that extrapolating ages from calcretes "is not justified."⁵
2. Early soil formation is fast, and the textbook says so. Retallack: "Organic surface horizons form in only tens to hundreds of years."⁶ A loess bank thrown up by railroad construction a hundred years earlier had gained 1.7% clay.⁶ Michael Oard is right that pedogenesis starts quickly. The fastest figure anywhere in these sources is decades.
3. The Columbia River Basalt tension is real and was not invented by creationists. Kevin Hobbs and Judith Parrish reported in the GSA Bulletin in 2016 that the paleosols between the flows show "relatively long duration of pedogenesis (10⁴ to >10⁶ yr)," while some argon dates compress all the Grande Ronde lavas into 0.42 million years.⁷⁸ Two mainstream clocks disagree by a large factor. Neither is near a year: the shorter still needs 420,000 years for the lava alone.
4. The load-bearing young-earth argument is that paleosols lack an A horizon, and the book it cites explains the observation. Oard: "with few exceptions paleosols lack a top organic horizon."⁸ Klevberg, Oard and Bandy: "complete absence of the A horizon makes a paleosol claim suspect."⁹ Retallack's Soils of the Past, which they cite repeatedly, gives the mechanism and the timescale: burial gleization below the water table, aerobic decomposition above it, "a few hundred years," while "horizonation and root traces persist through these deep burial alterations well into greenschist facies of metamorphism."⁶⁶⁶ Missing organic carbon is what the mainstream model predicts.
5. A published field test separates a soil nodule from a nodule grown after burial, by observation rather than assumed rates. On the Late Ordovician Potters Mills paleosol in Pennsylvania, the calcrete nodules "truncate and are truncated by the burrow walls."³ A nodule that cuts a burrow and is cut by it was there while the burrow was open at the surface. Retallack lists five more discriminators of the same kind.⁶ All three young-earth alternatives, flood flocculation, hydrothermal alteration and diagenesis, need the nodule to form after burial, so all three fail the test.
6. The counts are verified at the primary and nobody disputes them. Retallack 1983, in the GSA Bulletin: "at least 87 successive fossil soils in 143 m of stratigraphic section" in Badlands National Park, which is 1.64 m per soil.¹ Klevberg, Oard and Bandy quote the figure with an exclamation mark; Oard quotes it again in 2017.⁹⁸ In more than forty years nobody on the young-earth side has published a re-examination of the section.
7. The arithmetic is not close. If the entire 371-day Flood were spent on those 143 m and nothing else, each soil would get 4.3 days. Retallack's ceiling for the very weakest class of profile, the class that still shows its parent bedding, is 40 to 80 years.⁶ The shortfall is a factor of about three thousand four hundred, and the real shortfall is worse, because the White River Group is a thin slice of a column the same year has to build. This days-per-soil calculation is this article's own, from Retallack's count and a 371-day Flood.
8. The Chinle Group is the cleanest test case, because no young-earth boundary puts it outside the Flood. Lawrence Tanner reports drab root traces, rhizocretions up to 30 cm, root traces up to 70 cm long, desiccation cracks 50 cm deep, and, in the Owl Rock Formation, a calcrete profile reaching Machette Stage IV with a laminated indurated cap.^10101010 Running his nodule diameters, 2 to 10 cm, through Retallack's own chronometer gives 5.0 to 8.6 thousand years per profile. Subtract the full stated standard error and one Chinle calcic soil still costs about 3,000 years. That chronometer is the weak one: n = 9, R² = 0.57, standard error 1.8 kyr. A second measurement, of carbonate mass rather than nodule size, gives a much larger number. Michael Machette weighed the carbonate: a Stage IV calcic soil at Las Cruces holds 129 grams of CaCO₃ in every square-centimetre column, and the fastest accumulation rate he measured anywhere in the Southwest is 0.51 g cm⁻² kyr⁻¹.¹¹ That is 253,000 years for one Chinle profile, computed in the direction most generous to a Flood.
9. The rhythm of the stacking is dated by a clock that owes nothing to soil science. In the Willwood Formation of Wyoming, "the rhythmic stacking pattern of the paleosols is driven by orbital precession cycles," and the age model agrees with deep-sea cores dated by other means.¹² No young-earth source found anywhere addresses this.
10. The young-earth literature has never examined a paleosol inside a Flood-deposited sequence. Its one field study, by Klevberg, Bandy and Oard in 2007, is real work at real outcrops, and every one of those outcrops is Pliocene to Pleistocene colluvium that their own model places after the Flood.¹³ Whitmore and Garner concede Morrison paleosols in print and cite nothing that examines one.¹⁴ Tim Clarey's Flood-model papers, whose boundary places the Chinle, the Morrison, the Willwood and the Badlands section inside the Flood year, do not contain the word.

The thing that decides it

Not the ages.  Those come from chronometers weak enough to attack, and this article attacks them first.  What decides it is geometry that carries no rate assumption:  a burrow wall cutting a carbonate nodule and being cut by it, a root trace better preserved inside a nodule than in the mud around it, an uncrushed turtle shell inside a nodule and a crushed one outside.³⁶  The hard parts of the bed existed while the surface was still open to burrowing animals.  Every young-earth alternative makes the hard parts after burial.

Part 1: What a paleosol is, and the test both sides accept

A soil is not a layer.  It is a profile:  a vertical sequence of horizons made from the top down by rain, roots, worms and time acting on whatever was lying there.  Bury it and the profile can survive.  What survives is a paleosol.

Retallack's Soils of the Past is the field's textbook,⁶ and it states the recognition criteria in one place:

"There are three main kinds of features by which paleosols may be recognized in the field and from laboratory studies:  root traces, soil horizons and soil structure.  Using these and other observations it is generally possible to distinguish paleosols from unaltered sedimentary deposits, volcanic flows or zones altered by faulting.  Paleosols can be altered by groundwater, by hydrothermal activity or by metamorphism, and in these cases often have a mix of features difficult to disentangle."⁶

Read the last sentence.  Retallack raises the hydrothermal, groundwater and metamorphic alternatives himself, on the page that states the criteria.  Klevberg's 2003 paper⁹ and Oard's 2017 article⁸ both present those alternatives as objections the field has not considered.

Root traces

A root trace tapers and branches downward.  A burrow does not.

"Unlike other trace fossils such as burrows, most root traces taper and branch downward.  They also are very irregular in width.  Large, near-vertical root traces characteristically have a concertina-like outline because of compaction of surrounding sediments. …  Despite these characteristics the distinction between root traces and burrows is not always easy."⁶

That last sentence is Retallack's own, in 1988 and again in 2001.¹⁵  He disposes of the gas-escape objection on the page where he raises it:  escape structures "are not so copiously branched or pervasive as root traces."⁶  And he states the limit plainly.  Root traces do not exist before the Silurian, and burrows in soils run back only to the Ordovician, so for older rocks neither is any use.⁶

The strong claim is that a single unambiguous root trace is enough:

"A gray shale with clear bedding planes may look like an ordinary sedimentary deposit but the existence of a few fossil root traces in growth position means that it was once a soil."⁶

Horizons and structure

Sedimentary layering has sharp boundaries because each layer was a separate delivery of sediment.  Soil horizons grade into each other, because a gradient of water and biology worked down through one mass of material.  Retallack in 1988:¹⁵

"Soil horizons usually have more gradational boundaries than seen in sedimentary layering. …  Compared to sedimentary layering, metamorphic foliation, and igneous crystalline textures, soil structure appears massive, hackly, and jointed."¹⁵

The young-earth side accepts the test

Tas Walker's 2003 reply to the geologist Joe Meert² gives the same three features:  "The three main field features used to interpret a paleosol are root traces, soil horizons, and soil structures."²  Retallack's words, in Retallack's order.  Walker then concedes the stakes:  "Clearly Meert considers that paleosols have the potential to refute the global Flood.  We agree!"²

John Whitmore and Paul Garner state the prediction sharply, in the 2008 International Conference on Creationism proceedings:¹⁴

"True paleosols (rank = 3).  Since the Flood involved the rapid accumulation of sediments without the passage of long periods between the deposition of individual layers, we would not expect the development of true soil horizons in Flood-deposited sediments."¹⁴

That is a falsifiable prediction, published by young-earth geologists, about a class of feature both sides agree on how to recognise.  The rest of this article is the test.

Part 2: Where the young-earth critique is right

Three of the objections are good ones, and are worth stating at full strength from mainstream sources before anything else.

The clocks are weak, and their authors say so

Two published equations convert a paleosol into a number of years.  Sheldon and Tabor's 2009 review⁴ reproduces both, then says this about them:

"While these chronofunctions hold some promise for paleosols with either Bt or Bk horizons, they are both based on relatively few data (n = 10 for Sheldon, 2003;  n = 9 for Retallack, 2005b) and many soil/paleosols types lack those particular horizons.  Semi-quantitative relationships discussed in Birkeland (1999) also offer some promise for at least making appropriate order-of-magnitude age estimates, but significantly more work is needed in this area."⁴

The clay-horizon equation has no error function at all, because "the relation falls apart for very long-formation times."  The carbonate-nodule equation has R² = 0.57 and a standard error of 1.8 thousand years.⁴  Nine points and an R² of 0.57 is a weak fit.  Anyone using that equation, this article included, is using a weak instrument.

Kraus's 1999 review⁵ is equally frank about the underlying assumption:

"The Leeder (1975) and Retallack (1983, 1984) approach assumes that times of formation of modern soil analogs can be reliably determined.  The pitfalls of this assumption for calcretes have been thoroughly discussed by Wright (1990).  He concluded that this approach is not justified because Quaternary calcretes actually show considerable variability in their rates of formation depending on local rainfall and the availability of airborne carbonate dust."⁵

She adds Retallack's own criticism of the method she helped build:  reading all the time in a section off the paleosols in it "is probably an unrealistic assumption."⁵

Soil starts fast

Klevberg and Bandy's companion paper¹⁶ argues that soil development is a sigmoidal function of time with a steep start, so long-age estimates come from wrongly assuming a straight line.¹⁶  On the early part of the curve the textbook agrees:

"Organic surface horizons form in only tens to hundreds of years, but it takes millions of years to form a horizon as deeply weathered as an oxic horizon."⁶

Retallack's fastest measured clay accumulation is a hundred-year-old railroad cut:

"One of the most rapidly formed clayey subsurface horizons documented is the accumulation of 1.7% additional clay in soil on porous loess thrown up only a hundred years ago by railroad construction"⁶

He also concedes that some properties stop changing early:  profile structure and melanization reached steady state within 10,000 years on the Merced River chronosequence, while others had not after half a million.⁶

So the first stages of a soil are cheap in time.  What no source here supplies is a stage cheap in days.  The cheapest published figure is the hundred-year railroad cut, and it bought 1.7% clay.

The Columbia River Basalt problem

This is the strongest single point in Oard's 2017 article, and he did not invent it.  He quotes Kevin Hobbs and Judith Parrish, in the GSA Bulletin in 2016, stating the tension themselves.⁷  Their own abstract puts the soil side of it in one clause:  the paleosols between the flows show "relatively long duration of pedogenesis (10⁴ to >10⁶ yr)."⁷  The passage Oard quotes puts the lava side against it:

"At least 50% of the Grand Ronde Formation (the R2 and N2 units) was erupted over a period of not more than 0.30 m.y. (Baksi, 1989), and some evidence suggests that all Grand Ronde lavas were erupted in just 0.42 … m.y. (Barry et al., 2010).  Both of these proposed eruptive durations present interesting problems for the correlation between eruption timing and pedogenesis, because many paleosols within these units are interpreted to represent hundreds of thousands of years of landscape quiescence during which pedogenesis occurred."⁸

Two mainstream clocks, argon dating of lava and soil development on the interbeds, are hard to reconcile.  That is a genuine defect in somebody's method, and a reader is entitled to be told about it.

The p. 1545 passage could not be read directly:  the full text is closed at GeoScienceWorld and no repository copy exists.  Oard's transcription is not exact in every detail, because he prints "Grand Ronde" where the paper's abstract spells the unit "Grande Ronde."  Nothing turns on that, because the abstract carries the same tension alone.

It is also worth saying how far it gets a Flood model.  The dispute is between roughly 0.42 million years and something larger.  A Flood year is 0.0000027 million years.  If the soil clock runs fast by a factor of ten, the paleosols still cost tens of thousands of years each;  if the argon clock runs slow, the answer moves the wrong way for a young earth.  Two long chronologies disagreeing does not make either of them short.

Part 3: The missing A horizon, answered in the book that is cited against it

This is the argument the whole young-earth position rests on.  Oard states it in its baldest form:

"All soils have soil horizons, but with few exceptions paleosols lack a top organic horizon, one of the best indications of a buried soil.  The claimed paleosol is most often only a clay layer, carbonate layer, a 'weathered layer', or a mudstone, which can easily be a depositional layer or a layer altered by diagenetic or hydrothermal reactions."⁸

Klevberg, Oard and Bandy put it the same way in CRSQ in 2003:  the truncated top "usually means that the organic enriched A horizon is missing … leaving behind a clay layer, a calcium carbonate layer, a red layer, etc. …  However, it is the A horizon that would most readily identify a layer as a paleosol. …  complete absence of the A horizon makes a paleosol claim suspect."⁹

The observation is true.  Most paleosols have lost their dark organic top.  The question is whether that is an anomaly for the mainstream model or a prediction of it.

Chapter 7 of Soils of the Past is about what burial does to a soil, and Klevberg, Oard and Bandy cite that book repeatedly, including for the observation itself.

Below the water table, bacteria eat it:

"Drab horizons and mottles also can form by burial gleization … This is the chemical reduction of iron hydroxides and oxides by anaerobic bacteria consuming organic matter buried with the soil below or near the water table.  Reduction haloes around buried organic matter can form in only a few hundred years"⁶

Above it, oxygen eats it:

"Such observations support the idea that subsidence below water table is necessary for active microbial reduction of buried organic matter.  In contrast, paleosols buried and remaining above water table lost organic matter by aerobic decomposition."⁶

And here is what survives:

"Burial decomposition, burial gleization, burial reddening and cementation can significantly alter the appearance of paleosols very shortly after burial. …  Many physical features of paleosols such as horizonation and root traces persist through these deep burial alterations well into greenschist facies of metamorphism.  It is not until the amphibolite metamorphic facies that physical structure is completely obscured by recrystallization."⁶

Retallack said the same thing in the abstract of the 1988 chapter Klevberg cites:  "In oxidized paleosols their organic matter may not be preserved, but root traces can be recognized by their irregular, tubular shape, and by their downward tapering and branching."¹⁵

So the mainstream model says:  organic carbon is destroyed within a few hundred years of burial by one of two ordinary processes, while the physical structure survives into low-grade metamorphism.  Root traces and gradational horizons in a bed with no organic carbon are exactly what it predicts.  A young-earth reply has to argue that the mechanism is wrong;  it cannot argue that the observation is unexplained, because the explanation is in the book the argument cites.  No such reply was found in the young-earth literature searched here.

Two other objections that are answered on the pages they cite

Relict bedding.  Klevberg, Oard and Bandy write that "One should especially be suspicious when the supposed soil-forming process was not enough to wipe out bedding planes or even ripple marks."⁹  But surviving bedding is the definition of the weakest development class, not a refutation of it:  "Little evidence of soil development apart from root traces:  sedimentary, metamorphic or igneous textures remain from parent material."⁶  Retallack tracks its disappearance from "much relict bedding remains" through "some traces … within peds" to "no trace of relict bedding."⁶  Kraus's Willwood figure caption shows both ends of that scale in one outcrop.⁵

Diagenetic mimicry.  Klevberg, Oard and Bandy cite Kraus for the claim that diagenesis can mimic soil features.⁹  Kraus does say that:  in Paleogene rocks in Portugal, "Pimentel et al. (1996) found early diagenetic groundwater alteration features that mimic hydromorphic soil features such as pseudogley mottling and calcic horizons."  Her next sentence, which the citation omits, says the mimicry can be detected:  "Pimentel et al. provided criteria for distinguishing between the diagenetic and pedogenic features;  however, they cautioned that the diagenetic features could be mistaken for pedogenic features and lead to incorrect paleoclimatic interpretations."⁵  That is a real caution and it is quoted here whole, because the second half of it helps Klevberg.  What it does not say is that the two cannot be told apart.  Kraus's point in this section is narrower still:  these particular Portuguese features carry no climate signal.  The problem is cited;  the published test for it is dropped.

Part 4: The nodule test, which needs no clock

Suppose every rate argument above is wrong.  A class of evidence still decides the question, because it is geometry rather than arithmetic.

The three young-earth alternatives to a paleosol are all post-burial processes.  Klevberg, Oard and Bandy list them:  rapid flood deposition, where "a clay layer can be deposited rapidly by flocculation, while a red layer can be the result of iron-rich water";  hydrothermal alteration;  and diagenesis, because "Diagenetic effects can easily mimic paleosols."⁹  Oard adds the general form:  "During the Flood, we would expect rocks to have the appearance of weathering since rapid sedimentation and fluid flow would be common."⁸

Each of those makes the hard parts of the bed after burial.  So any observation showing that a hard part existed while the surface was still open is fatal to all three at once.

Mutual truncation.  On the Potters Mills clay paleosol in the Late Ordovician Juniata Formation of central Pennsylvania, a surface network of 1 to 15 mm burrows sits above a horizon of calcrete nodules of about the same size.  Retallack, in a 1992 survey of Paleozoic paleosols:³

"The nodules truncate and are truncated by the burrow walls indicating that the burrows are part of the original soil."³

A nodule cannot be cut by a burrow unless it was there when the animal dug.  A burrow cannot be cut by a nodule unless the nodule kept growing afterwards.  Both relationships in one outcrop put the nodule and the open burrow in the same episode, at the surface.  Note what this profile is:  Ordovician, older than land plants, no root traces available.  The recognition still works, on burrows and nodules alone.

Roots avoiding, entering and being protected by nodules.  Retallack's list of the field relationships that discriminate a soil nodule from a burial nodule:

"Sand-filled burrows approaching, then sidling around siderite nodules have been found …, unequivocally indicating that the nodule or some precursor of it was there when the burrows were excavated. …  Fossil root traces may sidle around them, but if penetrating the nodule are better preserved and more fully inflated than in the surrounding matrix.  Other fossils such as skulls, turtle shells and hollow bones also may be less compacted within nodules than outside them.  This is evidence that the nodules formed before there was any appreciable compaction …  Original pedogenic nodules also may contain more easily weatherable minerals such as pyroxene and olivine than are found in the surrounding matrix."⁶

Five observations, none of which requires knowing how old anything is.  An uncrushed turtle shell inside a nodule and a flattened one outside says the nodule was hard before the sediment compacted.  Pyroxene and olivine surviving inside and gone outside says the nodule sealed itself off from the weathering that stripped the soil around it.

The principle behind all of them is one line:  "Careful attention to relationships with unquestionably original structures, such as fossil root traces and burrows, allows discrimination between original features of a paleosol and those formed during burial."⁶

No young-earth source found addresses any of these.  What a reply would have to show is that rapid deposition plus early diagenesis can produce mutual truncation between a burrow wall and a nodule, or an uncrushed skull inside a nodule in a bed that was buried in hours.  Nothing of the kind was found.

Part 5: How many, and how long

Badlands: 87 soils in 143 metres

Retallack's 1983 GSA Bulletin paper¹ summarises his Badlands monograph, and its abstract carries the number everyone quotes, along with the criteria he used:

"Fossil soils can be recognized from irregular vertical tubular structures (fossil root traces and burrows), massive bioturbated layers with gradational contacts (soil horizons), complex cracking and veining (peds and cutans), and distinctive petrographic textures (sepic plasmic fabrics). …  In the Pinnacles area of Badlands National Park, there are at least 87 successive fossil soils in 143 m of stratigraphic section."¹

Klevberg, Oard and Bandy reproduce it:  "Retallack (1983) even managed to discover eighty-seven paleosols in a 143-meter stratigraphic section in Badlands National Park, South Dakota!"⁹  Oard quotes it again in 2017.⁸  Neither disputes the count.  Neither has visited the section and published what they found.

Now the arithmetic, which is this article's own and uses two verified inputs.

• 143 m of section, 87 soils. That is 1.64 m per soil.
• Genesis 7 and 8 give the Flood about 371 days. Give the whole of it to this one 143 m, which no Flood model actually can, because the same year has to build the rest of the Phanerozoic column. Each soil then gets 4.3 days.
• Retallack's ceiling for a very weakly developed profile, the class defined by still showing its parent bedding, is 40 to 80 years.⁶

40 years against 4.3 days is a shortfall of about 3,400 times, on an assumption maximally generous to the Flood model.  Restrict the White River Group to its real share of a one-year column and the figure per soil drops from days to hours.  And 4.3 days has to cover making the soil, letting plants root through it, and burying it under the next 1.64 m without washing the last one away, 87 times running.

Retallack's development scale, stated for peat and said by him to be "of the same order" for calcareous and clayey subsurface horizons, is the yardstick:

Those come from the standard textbook, not from a source hostile to the Flood.⁶

The Chinle: Stage IV calcretes inside everybody's Flood

The Badlands sit in the Oligocene, and some young-earth geologists put the Flood boundary below that.  The Chinle Group has no such escape.  It is Upper Triassic, and no published young-earth boundary puts it above the Flood.

Lawrence Tanner surveyed pedogenic features through the Chinle across the Four Corners in 2003.¹⁰  In the Blue Mesa Member he reports beds up to 8 m thick with "arcuate, intersecting slickensides, downward-tapering sandstone-filled fissures, drab root traces up to 0.1 m long," and "subspherical calcareous nodules that are 2–8 cm in diameter."¹⁰  In strata equivalent to the Rock Point Formation north of Durango, Colorado, "desiccation cracks that extend up to 50 cm deep, root traces up to 70 cm long, crumb and blocky mudstone fabrics, rhizocretions up to 30 cm long, and calcareous nodules up to 10 cm in diameter."¹⁰  Those Rock Point profiles are Stage II to III.  The top of the maturity scale is reached higher up, in the Owl Rock Formation:

"The mudstone-hosted calcareous nodule horizons represent simple Stage II to IV calcrete profiles (Gile et al., 1966;  Machette, 1985), or Calcisols (Mack et al., 1993).  Stage II profiles occur lower in the formation than the more mature Stage III and IV [profiles]"¹⁰

A Stage IV calcrete is not a clay layer that might be mistaken for a soil.  It is a plugged horizon capped by a laminated crust, built by carbonate carried down from the surface again and again until the pore space is full and water can no longer get through.  Tanner's photograph caption for the Owl Rock example, in the southern Echo Cliffs:  "Mature (Stage IV) calcrete profile in the lower Owl Rock Formation, southern Echo Cliffs.  A platy-laminar fabric caps the sequence and nodule density decreases downward.  The middle of the section displays a crude prismatic fabric."¹⁰

Put those nodule sizes through Retallack's published chronometer, age in thousands of years = 3.92 × (diameter in cm)^{sheldon-2009:chrono][}or print it with its own warning label:  n = 9, R² = 0.57, standard error 1.8 thousand years.⁴⁴

Subtract the full stated standard error of 1.8 kyr from the smallest and one Chinle calcic profile still costs about 3,000 years.  There are many of them, stacked, in a formation every young-earth model deposits during the Flood year.  This computation is this article's own, from Tanner's measured diameters and Retallack's equation.  The equation is weak, as Part 2 says;  it would have to be wrong by five orders of magnitude to fit.

The same Chinle calcretes, priced by weight instead of by nodule size

The nodule chronometer is the weak clock, and a critic is right to say so.  There is a second way to price a calcrete.  Weigh the carbonate.

Michael Machette, of the US Geological Survey, set out the calcrete stage scale Tanner is using, extending Gile and others' four stages to six.¹¹  He also measured how much secondary carbonate a soil holds, in grams of CaCO₃ per square-centimetre column through the profile, which he writes cS, on soils of known age in four regions of the American Southwest.  His Table 3 lists what 500,000 years buys in each:

His running text repeats the table:  "The Mescalero has accumulated carbonate at an average rate of 0.51 ± 0.06 g/cm2/1,000 years for the past 500,000 years."¹¹¹¹  Roswell is the fastest of the four, and Machette calls such rates "the highest in the Southwest."¹¹

Now run the sum in the direction most generous to a Flood model.  Take the Stage IV profile Machette measured at Las Cruces, 129 g/cm², and build it at the fastest rate he measured anywhere, the 0.51 g cm⁻² kyr⁻¹ at Roswell.  It needs 253,000 years.  Do the same for the cheapest Stage III profile, 71 g/cm² at Beaver:  139,000 years.  Tanner's Owl Rock profile is Stage IV.  This computation is this article's own, from Machette's Table 3.  It is a floor only in the sense that no rate measured in the Southwest is faster.  A wetter Triassic with more airborne carbonate could in principle run faster;  Machette measured no such case, and his four regions differ by a factor of under four.

Two measurements of the same profile disagree by a factor of about thirty:  8.6 thousand years from nodule diameter, 253 thousand from carbonate mass.  Both are far outside a Flood year.  But be clear what the second one is.  Machette's rates are a measured carbonate mass divided by an age taken from the standard Quaternary chronology, so this is a consistency check, not a clock independent of that chronology.  What it shows is that a Chinle Stage IV profile holds about as much pedogenic carbonate as soils the Southwest chronology places at half a million years.  A young-earth reply has to speed up the carbonate supply itself, not merely reject the dates.

Retallack's textbook repeats the same rates from the same sites:  0.26 g cm⁻² kyr⁻¹ at Las Cruces, "about 0.09 g cm⁻² kyr⁻¹" near Boulder, Colorado, and 0.51 at Roswell.⁶  His conclusion from that spread:  "natural variation in rates is less than the several orders of magnitude in time spans."⁶  Climate moves the rate by a factor of about six, not by a factor of a million.

The Willwood: the stacking has a period

The Willwood Formation of Wyoming is the most intensively studied paleosol stack in the world.  Bown and Kraus sorted its floodplain paleosols into seven development stages across a composite section running from the base to the 600 m level.⁵  Retallack notes that one vertebrate sample there may come from up to four paleosols, "which together may represent 50 kyr or so of soil formation."⁶

The part with no young-earth reply is the rhythm.  The Bighorn Basin Coring Project¹² tied the outcrop paleosol stack to drill cores and then to the deep sea:

"The rhythmic stacking pattern of the paleosols is driven by orbital precession cycles and was successfully used to establish a cyclostratigraphy for the PETM at PCB … roughly consistent with independent age models from deep-sea cores"¹²

A Flood model has to explain why hundreds of soil-like beds exist in a fluvial formation.  It then has to explain why they alternate on a period matching an astronomical cycle, and why that period agrees with marine sections on the other side of the planet.  The obvious alternative is autocyclic river avulsion, and Kraus gives its timescale:  avulsion cuts sedimentation off from part of a floodplain "for periods of time that are probably on the order of" a thousand years.⁵  That is the fast, self-organising, no-climate-needed explanation, and it runs a thousand years per cycle.

Why this is not a rare curiosity

The problem is not one section in South Dakota.  Kraus, summarising the field:  "It now appears that soil development, due to the episodic nature of sediment accumulation, is a normal part of the continental sedimentary regime and that many ancient continental deposits contain vertically stacked or multistory paleosols. …  Their geologic history is also lengthy, extending well back into the Precambrian"⁵

Part 6: What the young-earth literature has and has not done

The one field study, and where it is

Klevberg, Bandy and Oard published a two-part field investigation in CRSQ in 2007.¹³  It is the only creationist field work on paleosols in print and it deserves credit:  they went to outcrops on the Flaxville Plain in Montana and Alberta, took samples, and argued from what they saw.  Their result:

"Evidence of ongoing soil formation was observed at all of the sites described by EGP researchers, as was evidence of slope instability.  These sites do not reveal a cross section of superposed soil profiles, but they show soil formation within active colluvium."¹³

Take that at face value.  It still does not reach the question.  Those sites are Pliocene to Pleistocene surficial deposits on an erosion surface, which the authors' own model places after the Flood.  Showing that Montana colluvium is not a stack of interglacial soils says nothing about the Badlands, the Willwood or the Chinle, and no young-earth field study of any of those was found.

The one quantitative model, and what it costs

Klevberg, Oard and Bandy do price a stacked sequence, once:

"For example, suppose an EGP researcher believes ninety paleosols are represented in the wall of a canyon.  Although unable to disprove the paleosol interpretation, field evidence may suggest that two-thirds of the 'paleosols' are better explained by energetic geologic processes.  If the average time required for a soil in the paleopedologic sequence to form was 150 years (and negligible time for the geologic events), the resulting total time for formation of the sequence would have been 4,500 years."⁹

Their own example throws out two-thirds of the soils, prices the rest at 150 years each, and still needs 4,500 years:  the entire post-Flood period on a young-earth chronology.  And they set it in a canyon wall, inside a Flood sequence, where the model has no years to spend.  The paper does not say where the 4,500 years fit, and the 150-year figure is asserted:  no chronosequence, calcrete stage or measured rate appears anywhere in it.

The silences

Tim Clarey's Flood model contains these formations and never mentions their soils.  His megasequence framework runs the Flood from the base of the Sauk to near the top of the Cenozoic, placing the Chinle, the Morrison, the Willwood and the Badlands section inside the Flood year.  A search for "paleosol" across nineteen Clarey papers in this project's source library returns nothing in any of his Flood-model papers, the three megasequence papers included.¹⁷  The one file that uses the word is his 2016 piece on the Glasgow Fossil Grove, where he argues for a paleosol on the mainstream criteria.  A silence is not a refuted claim, but an author who assigns a formation to the Flood year owes an account of the soils in it.

Whitmore and Garner concede the Morrison and cite nothing that examines it.  Their 2008 paper:  "Paleosols are claimed at various horizons within the Morrison (Demko, Currie, & Nicoll, 2004), as well as apparently in situ termite nests in the Salt Wash Member (Hasiotis, 2004)."¹⁴  They rank paleosols only a tertiary criterion "because of uncertainties concerning their recognition and true interpretation," and cite against them only Oard 1990, a book about Pleistocene interglacial soils, and Walker 2003, a web reply.¹⁴  Neither examines a Mesozoic soil.  Andrew Snelling's 2018 boundary paper restates the criterion and adds nothing.¹⁸

A CRSQ paleosol series by Carl Froede does not exist.  A title search of the CRSQ archive returns exactly four papers with "paleosol" in the title:  Klevberg, Oard and Bandy 2003, Klevberg, Bandy and Oard 2007 Parts I and II, and a 1995 paper by Williams and colleagues on Big Bend fossil wood.  None is by Froede, and an author search on Froede returns twenty CRSQ papers, none about paleosols.  His treatment is in his 1998 book Field Studies in Flood Geology, pp. 23 to 28, which Klevberg cites five times and which could not be obtained.  Nothing here rests on it.

Three citation defects, reported plainly

None of these changes anybody's conclusion.  All three are checkable.

1. Klevberg, Oard and Bandy 2003, p. 137 cite "Retallack (1988, p. 21, 22)" for the claim that root traces alone are diagnostic.⁹ Retallack's 1988 chapter runs from p. 1 to p. 20. There is no p. 21 or 22 in it. The same sentence then cites "(Retallack, 1990, p. 21, 22)" for the counter-claim, identical page numbers for a different work, so the likeliest reading is a slip for the 1990 book. The substance is right and the locator is wrong: Retallack does say it, on p. 2 of that chapter and p. 13 of the 2001 textbook.¹⁵⁶
2. The same paper, p. 141 cites "(Kraus, 1999, p. 61)" for diagenetic mimicry. The passage is on p. 63; page 61 is about Carboniferous cyclothem paleosols and a rainfall proxy. The next sentence, which the citation omits, both gives criteria for telling the two apart and cautions that they can still be confused.
3. Oard 2017 gives Retallack's Badlands monograph as "GSA Special Paper 216." It is Special Paper 193, printed on the face of the 1983 GSA Bulletin summary: "this article represents a summary and introduction to the Geological Society of America Special Paper 193."¹ Special Paper 216 is the 1988 Reinhardt and Sigleo volume containing the "Field recognition of paleosols" chapter. The 2017 article merged the two; Klevberg, Oard and Bandy cite both correctly in 2003.

Where this sits in the young-earth boundary dispute

Whitmore and Garner end the Flood in Wyoming above the Lance Formation, because "no more marine deposits occur" above it.¹⁴  On that placement the Eocene Willwood is post-Flood and its paleosols cost them nothing.  Clarey and the Institute for Creation Research run the Flood to near the top of the Cenozoic, so the Willwood is inside it.  The Flood/post-Flood boundary article covers that dispute.

Whichever placement is chosen, the Chinle and the Morrison stay inside the Flood, because every published young-earth boundary sits above the Triassic.  The disagreement decides only whether the Willwood is a second problem on top of the first one.

Objections and responses

"No A horizon means it is not a soil."  The strongest young-earth objection, and it has a published answer.  Organic carbon is destroyed after burial in a few hundred years, by anaerobic bacteria below the water table and by oxidation above it, while horizonation and root traces survive to greenschist facies.⁶⁶⁶  A buried soil that still had its dark organic top would be the anomaly.  To keep the objection alive someone has to show the mechanism is wrong, and nobody has tried.

"The nodules are diagenetic, or hydrothermal."  Both processes work after burial.  Neither can make a nodule that a surface burrow cuts and that cuts the burrow back.³  Neither can leave an uncrushed turtle shell inside a nodule and a flattened one outside.⁶  Retallack raises the hydrothermal alternative himself on the page that states the recognition criteria, and gives the tests.⁶  The objection is answered in the source it is aimed at.

"Root traces are just plant debris the Flood buried."  Klevberg, Oard and Bandy:  "One would expect all kinds of roots and root traces in sediments or sedimentary rocks caused by the Deluge, since the Flood-borne sediments would have buried many."⁹  Buried debris and a root system in growth position look different, and the difference was described before the objection was raised:  growth-position traces taper and branch downward, are irregular in width, and compact into a concertina outline.⁶  Tanner's Chinle rhizocretions, up to 30 cm and mineralised around the root, are not floating debris.¹⁰  And the objection cannot reach the Potters Mills profile at all, which is Ordovician and has no roots in it.

"Calcretes form fast in a warm, wet post-Flood climate."  Climate does change the rate, and the measured spread is the answer:  across a real climatic range the rate varies by a factor of about six, and "natural variation in rates is less than the several orders of magnitude in time spans."⁶⁶  A Flood year needs six orders of magnitude, not one.  There is a second problem for this reply:  pedogenic calcite forms in dry, well-drained soils under low rainfall and high evapotranspiration.⁴  A warm wet post-Flood world is the wrong climate for making calcrete.

"The Columbia River Basalt shows the soil clock is broken anyway."  It shows two mainstream clocks disagreeing, and Hobbs and Parrish say so in print.⁸  Take the young-earth side in full:  suppose the interbed paleosols are ten times younger than the mainstream reading.  Tens of thousands of years each remain, and the lavas below them still take 420,000 years by the shortest published estimate.

"A paleosol is an interpretation, not an observation."  Walker:  "unless it has been historically attested, the concept of a paleosol is merely an interpretation."²  Klevberg and Bandy:  "since paleosols are historic by definition, they are not essentially scientific."¹⁶  This proves too much.  Applied consistently it disqualifies every Flood deposit and every megasequence, none of which was historically attested either.  Nor is it how these authors argue elsewhere:  Klevberg's 2007 field study reads unattested Montana sediments as colluvium on physical evidence, which is the same move.

"Soil scientists have described these for a century.  Why is this still argued?"  Two honest reasons.  First, the young-earth objections are not silly.  Early pedogenesis really is fast, most paleosols really have lost their A horizons, and the two published chronometers really do rest on nine and ten data points.  Second, the two literatures barely touch.  The mainstream field answered the missing-A-horizon problem in a textbook chapter and the diagenetic-nodule problem in a 1992 book chapter, and the young-earth literature cites both books without engaging either answer.

"You used a chronometer you just called weak."  Yes, and it is labelled every time it appears.  It would have to be wrong by five orders of magnitude to fit a Flood year.  It is also not the only measurement:  Machette's carbonate mass gives a number thirty times larger for the same Chinle profile, though his rates are themselves calibrated on the standard chronology.¹¹  And nothing depends on either.  Strike every rate from this article and the mutual-truncation geometry still stands.

Issues and unanswered questions

• Machette 1985 was read in fragments, not in full. No open route supplies the whole chapter. Table 3 and the sentences quoted from pp. 16, 17 and 19 come from Google Books snippets on a volume with no preview, and pp. 1 and 5 from a US Department of Energy scan. The numbers self-check against each other and against Retallack's textbook, but nobody here has read the surrounding argument, and the calibration soils are Quaternary alluvium in a desert, not Triassic floodplains.
• Abdul Aziz et al. 2008 in Geology, which counts the Willwood paleosols and ties them to precession, was not obtained; GeoScienceWorld returns a challenge page. The precession claim here comes instead from Westerhold et al. 2018, which is open access and cites it. Abels et al. 2013 was read in abstract only and is not quoted.
• Demko, Currie and Nicoll 2004, the Morrison paleosol paper Whitmore and Garner cite, was not obtained, so the Morrison is discussed here only through the young-earth concession. Bown and Kraus 1987, the Willwood pedofacies pair, reaches this article through Kraus 1999. Froede's Field Studies in Flood Geology 1998, pp. 23 to 28, is the one young-earth treatment of paleosols this article could not read.
• The Soils of the Past quotations were read through archive.org's search-inside on a print-disabled copy, with the leaf-to-page map taken from the item's own scan data. That OCR destroys superscripts. The units on the carbonate accumulation rate were recovered from a second, independent scan of the 1990 first edition and from Machette 1985, which prints them without superscripts; no exponent in this article rests on a guess.
• Related evidence sits in sibling articles: Retallack's Specimen Ridge paleosols in Yellowstone's petrified forests, soils on flat contacts in flat gaps, the rotted regolith under the Tapeats in the Great Unconformity, and root-penetrated underclays in coal and the Flood.

Verdict

Buried soils are a real and unanswered problem for a one-year Flood, and the young-earth literature has not engaged the evidence that makes them one.

Both sides use the same three criteria.  Whitmore and Garner predicted in print that true soil horizons would not be found in Flood-deposited sediment.  They are found, in quantity, in formations every young-earth boundary places inside the Flood:  a Stage IV calcrete in the Triassic Owl Rock Formation and 70 cm root traces lower in the same Chinle Group, 87 stacked profiles in 143 m at Badlands, hundreds in the Willwood stacked on an astronomical period.

The rate arguments are the weaker half of the case and are presented as such.  The stronger half is geometry.  A carbonate nodule that cuts an open burrow and is cut by it was in the ground while the ground was still a surface, and Flood deposition, hydrothermal alteration and diagenesis all act after burial.  Until a young-earth field study takes that relationship to an outcrop in the Chinle or the Badlands and shows it can be made in a week, the honest statement is that this evidence stands against a one-year Flood.

Sources